1. Field of the Invention
This invention relates to the use of thermostatic HVAC controls that are connected to a computer network. More specifically, the present invention pertains to application of specific adaptive waveforms to the setpoints programmed into thermostats in order to reduce energy consumption with out a subjective loss of comfort.
2. Background
Heating and cooling systems for buildings (heating, ventilation and cooling, or HVAC systems) have been controlled for decades by thermostats. At the most basic level, a thermostat includes a means to allow a user to set a desired temperature, a means to sense actual temperature, and a means to signal the heating and/or cooling devices to turn on or off in order to try to change the actual temperature to equal the desired temperature. The most basic versions of thermostats use components such as a coiled bi-metallic spring to measure actual temperature and a mercury switch that opens or completes a circuit when the spring coils or uncoils with temperature changes. More recently, electronic digital thermostats have become prevalent. These thermostats use solid-state devices such as thermistors or thermal diodes to measure temperature, and microprocessor-based circuitry to control the switch and to store and operate based upon user-determined protocols for temperature vs. time.
These programmable thermostats generally offer a very restrictive user interface, limited by the cost of the devices, the limited real estate of the small wall-mounted boxes, and the inability to take into account more than two variables: the desired temperature set by the user, and the ambient temperature sensed by the thermostat. Users can generally only set one series of commands per day, and in order to change one parameter (e.g., to change the late-night temperature) the user often has to cycle through several other parameters by repeatedly pressing one or two buttons.
Because the interface of programmable thermostats is so poor, the significant theoretical savings that are possible with them (sometimes cited as 25% of heating and cooling costs) are rarely realized. In practice, studies have found that more than 50% of users never program their thermostats at all. Significant percentages of the thermostats that are programmed are programmed sub-optimally, in part because, once programmed, people tend not to re-invest the time needed to change the settings very often.
A second problem with standard programmable thermostats is that they represent only a small evolutionary step beyond the first, purely mechanical thermostats. Like the first thermostats, they only have two input signals—ambient temperature and the preset desired temperature. The entire advance with programmable thermostats is that they can shift between multiple present temperatures at different times without real-time involvement of a human being.
Because most thermostats control HVAC systems that do not offer infinitely variable output, traditional thermostats are designed to permit the temperature as seen by the thermostat to vary above and below the setpoint to prevent the HVAC system from constantly and rapidly cycling on and off, which is inefficient and harmful to the HVAC system. The temperature range in which the thermostat allows the controlled environment to drift is known as both the dead zone and, more formally, the hysteresis zone. The hysteresis zone is frequently set at +/−1 degree Fahrenheit. Thus if the setpoint is 68 degrees, in the heating context the thermostat will allow the inside temperature to fall to 67 degrees before turning the heating system on, and will allow it to rise to 69 degrees before turning it off again.
Standard programmable thermostats are all designed with the same basic underlying premise: that the comfort of building occupants is maximized by maintaining a relatively constant temperature, at least for the duration of a given setpoint, and with the variations inherent in using a hysteresis band to trade comfort off against efficient operation and durability. That is, if a programmable thermostat has been programmed to maintain a temperature of 68 degrees Fahrenheit for 8 hours, it will cycle the HVAC system as needed to maintain that temperature.
However, academic research has shown that humans tend not to notice changes in temperature if (a) they are below a certain magnitude and (b) if the rate of change is sufficiently slow. For example a 1978 study found that people did not notice ramps less than 0.5° C./h (0.9° F./h). A 2004 study found that and that ramps up to 1.5° C./h (2.7° F./h) are unlikely to cause discomfort.
Because energy consumption is directly proportional to setpoint—that is, the further a given setpoint diverges from the balance point (the inside temperature assuming no HVAC activity) in a given house under given conditions, the higher energy consumption will be to maintain temperature at that setpoint), energy will be saved by any strategy that over a given time frame lowers the average heating setpoint or raises the cooling setpoint. It is therefore possible to save energy by adopting a strategy that takes advantage of human insensitivity to slow temperature ramping by incorporating a user's desired setpoint within the range of the ramp, but setting the average target temperature below the desired setpoint in the case of heating, and above it in the case of cooling. For example, a ramped summer setpoint that consisted of a repeated pattern of three phases of equal length set at 72° F., 73° F., and 74° F. would create an effective average setpoint of 73° F., but would generally be experienced by occupants as yielding equivalent comfort as in a room set at a constant 72° F. Energy savings resulting from this approach have been shown to be in the range of 4-6%.
It would be advantageous to create a temperature control system that would automatically generate optimized ramped setpoints that could save energy without compromising the comfort of the occupants. It would also be advantageous to create a temperature control system that could incorporate adaptive algorithms that could automatically determine when the ramped setpoints should not be applied due to a variety of exogenous conditions that make application of such ramped setpoints undesirable.